Twitter’s One Million Dollar Stylesheet
At a recent party conversation brought up iPhone applications and Twitter (ok, perhaps ‘party’ is used lightly here). Tweetie, Twitterific, TweetDeck, TwitBird (interestingly just labeled “Twitter” in the app store), Twittelator Pro and Echofon are all paid iPhone apps ranging from $1-$5 USD (and all have been in the top 20 paid social networking apps in the app store this week).
Their current sales combined are well over $1M USD with the possibility that Tweetie alone has passed that mark (this guess is based off of contrasting the top grossing app list with the top apps, taking apps that publish their data publicly to create a known data point). Interviews with other top app developers had guesses from $300,000 total to $200,000 a week for Tweetie, although the actual number is not disclosed and most likely not as sexy. Nevertheless, that is quite a bit of cash for access to Twitter on the iPhone, but just why so much?
Simple. There is no iPhone specific stylesheet. If you access twitter.com on your iPhone in Safari, you are redirected to m.twitter.com with a stripped down mobile version of Twitter.

Their current mobile version is screaming fast at less than 9KB with only four items loading (about a half a second on the ATT&T Edge network) which is really nice, except it lacks in common basic features such as follow and reply. It is optimized for a T9 keyboard layout, which is useless for the iPhone. If you go to a user you have to tap ‘Standard’ to then zoom in and tap follow.

This is why the apps are selling so well, because the user experience on the iPhone is remarkably bad. Which puzzles me. Why is it? You can detect the device and present a seperate stylesheet, which should take about an hour to make (I’m shocked a fan has not already done this) and helps to solve one of the major problems of using twitter. The lack of stylesheet means users will flock to devices, perhaps piling in another $1M in sales.
![Twitters One Million Dollar Stylesheet Photo Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_e.png?x-id=0f40f1a6-4aeb-43cc-8ace-45904c56c0d1)





Andrew,
I only agree partially. There’s an ongoing app vs web battle and ultimately we will likely see the web win again, but comparing e.g. Tweetie with a pimped, ajax-ified twitter.com web site is just not right.
There are lots of features, that make a real difference. Multi-account handling, drafts, caching, etc. to name only a few. And Tweetie is ultimately faster than using a Mobile Safari based front end, even if Twitter would try to optimize the page for mobile downloading.
The Twitter folks never really cared about the client part anyway. They’ve created a messaging infrastructure, added a lightweight API to it to transform it into a platform. In fact the Twitter.com website only recently got really usable. It’s been a mess in terms of usability for too long.
So Twitter simply doesn’t want to do clients, whether it’s native apps or optimized mobile web clients. They always kept constraining themselves to the basics and left the client-side innovation to third parties.
I think it’s a fantastic separation of concern.
Finally, I’m really happy to get atebits and other awesome iPhone developers my 3 US$. Tweetie is well worth two cups of coffee.
I would agree that they don’t compare but the mobile is so bad that an app is needed even for non power users that really get things out of Tweetie.
I’m really glad I’ve sent atebits my cash too.
Actually, this very bad implementation makes a lot of sense to me.
As you mentionned, the bandwith used with m.twitter.com is really light. When you download a client, the bandwith usage with this client is huge.
You said 9Ko for m.twitter.com? See what’s going on with tweetie: 25 Ko w/ XML, and let’s admit about 15 Ko with JSON. Also, you have to add the download of avatars: 4Ko (a timeline is made of 100 statuses, so you have to x100 in the worst case).
If you go to another country, your bill is going to explode with Tweetie, just because it’s full featured. If you are French like me, downloading a single Mo of data costs the end-user about €13.30 (it’s a shame btw), so by checking your timeline on tweetie, you already spend in the worst case €4 for a single connection.
This is why the mobile version of Twitter is the ONLY good alternative when you’re abroad and you want to keep up w/ your friends.
Hmm… bad implementation? That’s the way the twitter folks designed their API, or? And Tweetie actually caches avatars.
This would, however, qualify for an additional mode. File an enhancement request with atebits!
I didn’t mean “bad implementation”, I was referencing about the bad UX provided by m.twitter.com (ie last paragraph)
Agree with you on international use. If done well I would think the stylesheet option would be just as small and allow you to do basic tasks.
I’m no expert twitter watcher, but it seems to me that there are two entwined issues.
1. Having a bad mobile experience for iphone users has fostered an application ecosystem around twitter–important for any platform. This ecosystem then carries water for twitter, and does a far better job of innovating the UI than twitter could.
2. Twitter isn’t focused (at this time) on the end user experience–they are focused on the underlying platform. They prefer to let other companies innovate around UX while they deal with traffic and innovate around their core platform.
As you say, Andrew, doing a nice device specific stylesheet seems like it would be easy, but it would also ruin (or have a drastic effect on) all the software vendors who are investing in twitter apps (and cementing twitter’s status as the real time status platform).
I think we’ll see a Twitter app released soon which will hopefully gives us a powerful new tool.